Sometimes, a task bound to VideoSendStreamTest was called
after the underlying object had been destructed:
1. |test| goes out of scope.
2. There might still have been a task in fixture's queue,
setup by OnSendRtp (capturing [this]) and invoked before
the destruction of the fixture.
This CL uses the same workaround than BandwidthStatsTest:
block until all posted tasks are processed.
This fixes the following flaky tests:
* VideoSendStreamTest.ChangingNetworkRoute
* VideoSendStreamTest.RespectsMinTransmitBitrate*
Bug: webrtc:11156
Change-Id: I229c96d2abbbb60b43e9d9f62ad112507a21fe48
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/163984
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Commit-Queue: Yves Gerey <yvesg@google.com>
Cr-Commit-Position: refs/heads/master@{#30152}
This CL replaces the use of the ExperimentalAgc config with
using the new config format.
Beyond that, some further changes were made to how the analog
and digital AGCs are initialized/called. While these can be
made in a separate CL, I believe the code changes becomes more
clear by bundling those with the replacement of the
ExperimentalAgc config.
TBR: saza@webrtc.org
Bug: webrtc:5298
Change-Id: Ia19940f3abae048541e6716d0184b4caafc7d53e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/163986
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30149}
The IPAddress class (32 bytes) was copied for each invocation.
This CL also saves some bytes in generated binary.
Bug: webrtc:9855
Change-Id: I40f2fe8570ee30d1d2251fddd56131ca4c3e7155
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/164521
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Commit-Queue: Yves Gerey <yvesg@google.com>
Cr-Commit-Position: refs/heads/master@{#30147}
This change means tasks scheduled at the end time reached when making a call to GlobalSimulatedTimeController::AdvanceTime will also be executed.
In other words, with this change, if you schedule a task in X milliseconds and then call AdvanceTime(TimeDelta::ms(X)) the scheduled task will be executed.
Bug: none
Change-Id: I337e574a88b235639e82ffcacf1484daa6cf3172
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/164522
Reviewed-by: Sebastian Jansson <srte@webrtc.org>
Commit-Queue: Philip Eliasson <philipel@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30146}
There's a similar member in the base class, PortAllocator, which
appears to be in use.
Bug: None
Change-Id: Ie82801a7d0ae62f1e2758b6f434485bd5f78e8ba
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/164520
Reviewed-by: Harald Alvestrand <hta@webrtc.org>
Commit-Queue: Niels Moller <nisse@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30145}
The two functions have a lot of shared logic and locking. This CL consolidates that into a single function.
Bug: webrtc:111235
Change-Id: Ib1c32165dbf0e212c7d4b0753bcbb5ffd05eb6fe
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/163022
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Reviewed-by: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30144}
This CL ensures that the pre-amplifier and the high-pass filter
submodules are not reallocated more than needed.
Bug: webrtc:5298
Change-Id: I7ed23807d4d2d9fef0eda2e7dca9de9b0b1a4649
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/163988
Commit-Queue: Per Åhgren <peah@webrtc.org>
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30143}
Degradation preference could be changed before video send stream
is configured which would cause a crash.
Bug: None
Change-Id: If970e66fba0b9fdb9da789066861d919874de119
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/164463
Reviewed-by: Niels Moller <nisse@webrtc.org>
Commit-Queue: Mirta Dvornicic <mirtad@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30141}
This brings the two ProcessStream functions closer in implementation.
Additionally, the error checking that is currently done in the period of not holding the lock seems cheaper than releasing and reacquiring the capture lock.
Bug: webrtc:11235
Change-Id: Ib4afc68afb419fcabbb8cf08a3a2e61d2c12acda
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/163021
Reviewed-by: Per Åhgren <peah@webrtc.org>
Commit-Queue: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30140}
This CL ensures that the AGC2 is created and initialized only when
needed.
Apart from that, the CL also avoids a runtime-reallocation that happens
each time the setting is applied.
Bug: webrtc:5298
Change-Id: Iad9eaa05a3d0baa0788cd11b2aa17ddd8e0c509b
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/163987
Commit-Queue: Per Åhgren <peah@webrtc.org>
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30139}
This CL moves the activation of the transient suppression to the APM
config.
Bug: webrtc:5298
Change-Id: Iba7975bec4654c3df8834fd5b7d1082ff53641dd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/163985
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30137}
Replace all usages of java_files with sources in gn files, and
automatically format.
This is in preparation for java_files being completely removed upstream
in favor of sources.
NOPRESUBMIT=true
Bug: chromium:1035074
Change-Id: Ib9a698740b7ad26a127031d90321c7ae2feb59bf
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/163161
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Commit-Queue: Natalie Chouinard <chouinard@google.com>
Cr-Commit-Position: refs/heads/master@{#30135}
This CL changes the GetStatistics call in the audio processing module
(APM) to not aquire the render or capture locks in APM, while still
being thread-safe.
This change eliminates the risk of thread-priority inversion due to the
GetStatistics call.
Apart from the above the CL:
-Corrects the GetStatistics to not be const (it was const even though it
aquired locks).
-Slightly changes the statistics reporting, so that the stats received
may be older than the most recent stats reported.
Bug: webrtc:11241
Change-Id: I00deb5507e004cbe6e4a19a8bad357491f86f4ab
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/163982
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30131}
1. On iOS13 the implementation of methods begin- and endGeneratingDeviceOrientationNotifications changed and now are looks like "not threadsafe" (in specific sence) - they should be called only on the main thread. This fact is not documented. And may be a mistake.
2. By the Apple official documentation methods begin- and endGeneratingDeviceOrientationNotifications should be balanced. (Each begin- method should be balanced with end- method.)
By the reason two above facts they consequences merged and produced the "floating" NSInternalInconsistencyException crash.
Bug: webrtc:11216
Change-Id: Ibedd5bba7476cc687de3b9b04be49e3cceac1d27
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/162205
Reviewed-by: Anders Carlsson <andersc@webrtc.org>
Commit-Queue: Anders Carlsson <andersc@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30121}
This CL allows the noise suppressor to use the linear AEC output
for analysis whenever that is available. This will potentially
lower the risk that the noise suppressor estimates the noise
based on echo.
The feature is off by default.
Bug: webrtc:5298,b:132164318
Change-Id: Idc6c8e197d96209d213819d87a8fb2533b7303ec
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/162900
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30116}
This CL ensures that the no data vectors in the reverb computer code
are fixed. This allows arbitrary long filters to be used, and ensures
that a minimum required heap size is used.
Bug: webrtc:8671
Change-Id: I7085ed262a3f5965d796270434b6578f4030606e
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/162661
Reviewed-by: Sam Zackrisson <saza@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30115}
max-fr and max-fs are mandatory fields for VP8 and VP9.
Add parsing as a first step to enable use of these fields.
Bug: chromium:1032518
Change-Id: I4fd8f7f84f6303d646fb3f5313a02d6cf4160346
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/162801
Commit-Queue: Johannes Kron <kron@webrtc.org>
Reviewed-by: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30114}
This CL moves the high-pass filter to run in the full-band domain
instead of the split-band domain.
Bug: webrtc:11193
Change-Id: Ie61f4a80afda11236ecbb1ad544bbd0350c7bbfd
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/161453
Reviewed-by: Gustaf Ullberg <gustaf@webrtc.org>
Commit-Queue: Per Åhgren <peah@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30112}
The pacing controller allowed sending bitrate probes, despite it being
paused. This CL adresses that, and makes sure the task-queue based mode
also properly repsects pausing.
Bug: webrtc:10809
Change-Id: I79643c9a24666110d7583fce3ed1bfd6865e9e10
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/162520
Reviewed-by: Florent Castelli <orphis@webrtc.org>
Commit-Queue: Erik Språng <sprang@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30109}
If the first simulcast stream is disabled, but the second one is enabled,
FrameEncodeMetadataWriter would fail to store frame metadata for all
streams and later fail to restore it for encoded frames.
Bug: none
Change-Id: Ib0d257abb863716ea94e56730f7caabef6ebeb64
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/162480
Reviewed-by: Erik Språng <sprang@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30106}
If for some reason capture timestamp is unset, the default value of 0 would be
passed to RtcpSender. This will cause rtp timestamps to grow at double the rate
in Sender Reports because it has time since the last frame capture as a term.
Bug: none
Change-Id: I2fe09dabef6b0957fb504deaa06393dedc4a9e70
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/162481
Reviewed-by: Danil Chapovalov <danilchap@webrtc.org>
Commit-Queue: Ilya Nikolaevskiy <ilnik@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30105}
The network emulation framework does not support creation sockets that
receive from all addresses (e.g. 0.0.0.0) but would instead crash at
runtime. This CL explicitly ensures that we don't provide such networks.
Bug: webrtc:9883
Change-Id: I1d77df0f2c68f878eace30e4b037ebc7eb9f1aa6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/162482
Reviewed-by: Per Kjellander <perkj@webrtc.org>
Commit-Queue: Sebastian Jansson <srte@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#30104}
